Commercial mindfulness aid 2018

Received:
20 August 2018
Revised:
31 December 2018
Accepted:
12 March 2019
Cite as: Artem S. Svetlov,
Melanie M. Nelson,
Pavlo D. Antonenko,
Joseph P. H. McNamara,
Regina Bussing. Commercial
mindfulness aid does not aid
short-term stress reduction
compared to unassisted
relaxation.
Heliyon 5 (2019) e01351.
doi: 10.1016/j.heliyon.2019.
e01351
Commercial mindfulness aid
does not aid short-term stress
reduction compared to
unassisted relaxation
Artem S. Svetlov a, Melanie M. Nelson a, Pavlo D. Antonenko b,
Joseph P. H. McNamara a, Regina Bussing a,∗
a
Department of Psychiatry, College of Medicine, University of Florida, PO Box 100256, 1149 Newell, Dr., L4-100,
Gainesville, FL 32611, USA
b
∗
College of Education, University of Florida, G416 Norman Hall, PO Box 117042, Gainesville, FL 32611, USA
Corresponding author.
E-mail address: rbussing@ufl.edu (R. Bussing).
Abstract
Increased public interest in mindfulness has generated a burgeoning market in new
consumer technologies. Two exploratory studies examined effects of InteraXon’s
“Muse” electroencephalography (EEG)-based neurofeedback device and mobile
application on mindfulness-based relaxation activities. Psychophysiological
outcomes (heart rate variability (HRV), electro-dermal activity (EDA), saliva
amylase activity (sAA) and Muse application EEG “calm percent”) were
collected for two 7-minute conditions: Muse-assisted relaxation exercise
(MARE), and unassisted relaxation exercise (URE). In the first study,
participants (n ¼ 99) performed both conditions in a randomized sequential
design. A follow-up study used a randomized parallel condition (n ¼ 44) to test
for differences in HRV effects between the two conditions and extended followup observation. Generalized estimating equation models demonstrated a moderate
increase in HRV following relaxation exercises, with no observable difference
between MARE and URE conditions. Both MARE and URE conditions
produced equally effective short-term increases in heart rate variability, without
additional benefit from neurofeedback.
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Keyword: Psychology
1. Introduction
“Stress” refers to the subjective experience of sympathetic nervous system activation
in response to adverse stimuli. Situations that are novel, unpredictable, or appear to
be outside the subject’s control elicit this “fight, flight, or freeze” response. Repeated
or prolonged activation of the stress response has been associated with numerous
harmful effects (Brosschot et al., 2005; Salleh, 2008), including cardiovascular
and immune dysfunction, psychiatric disorders, and suicide risk. According to the
American Psychological Association’s 2017 Stress in AmericaÔ survey results (n
¼ 3,440 adults), nearly half of Americans reported sleeplessness due to stress
(45%), one-third of adults reported experiencing feeling nervous or anxious
(36%), irritability or anger (35%), and fatigue (34%) due to their stress (APA,
2017). Stress-related illness and injury are estimated to cost the United States
more than $300 billion per year, including costs related to stress-related accidents,
absenteeism, employee turnover, diminished productivity, and direct medical, legal,
and insurance costs (American Institute of Stress, 2013).
Physiologically, stress activates the body’s autonomic nervous system and the
hypothalamic-pituitary-adrenal (HPA) axis in the brain by releasing adrenaline,
corticotropin-releasing hormone, and cortisol. The degree of activation depends on
the severity of the stressor, but also the individual’s genetic background and life experiences. As individuals experience more symptoms of stress, they employ various
stress-management methods. For example, more than half of Americans (53%) exercise
or take part in physical activity to cope with stress, a significant increase in 2017
compared to each of the past three years (APA, 2017). The use of yoga and meditation
is rising, with 12 percent of people using these activities to manage their stress, the highest percentage since the survey first asked about these activities in 2008 (APA, 2017).
Mindfulness is an umbrella term that describes many self-awareness practices,
including some forms of meditation and yoga. While the term has its historical
footing in Buddhism (Bodhi, 2011), it has achieved popularity in psychology, psychiatry, medicine, neuroscience, and education through its central role in
mindfulness-based stress reduction (MBSR)da structured training program from
the University of Massachusetts Medical School researchers (Kabat-Zinn, 2003,
2011) The eight-week workshops incorporate weekly group instruction, daily individual practice, and a single day-long intensive to provide instruction in three formal
techniques: mindfulness meditation, body scanning and simple yoga postures.
A robust meta-analysis of meditation programs for psychological stress and wellbeing examined 47 randomized control trials with active controls for placebo effects
2
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
and a total of 3,515 participants (Goyal et al., 2014). Meditation programs had moderate evidence of reduced anxiety (d ¼ 0.38), depression (d ¼ 0.30), and pain (d ¼
0.33) at 8 weeks, resulting in moderate reductions of multiple negative dimensions
of psychological stress. A recent literature review did not find correlations between
mindfulness program length and effect size for psychological outcomes, concluding
that shorter programs using 1.5 h of instruction over 7 weeks may also be effective
for reducing distress (Carmody and Baer, 2009). A recent study suggests that daily
10-minute online MBSR interventions over two weeks may offer benefits for nonclinical groups seeking stress-reduction strategies (Cavanagh et al., 2013).
Regarding dose-response relationship between mindfulness practice and outcomes,
post hoc analyses of practice time or attendance and symptoms or self-reported measures yield inconsistent results, with some studies showing a dose-dependency
(Carmody and Baer, 2008; Collard et al., 2008; Farb et al., 2013; Rosenzweig
et al., 2010; Shapiro et al., 2012), but other not (Carlson et al., 2004; Dobkin and
Zhao, 2011; Hopkins and Proeve, 2013; MacCoon et al., 2012). The actual length
of the intervention might not matter as much as attendance and the quality of the
practice (DeVibe et al., 2012; Del Re et al., 2013). Several MBSR programs have
implemented online and mobile applications for treatment delivery to extend the
reach of mindfulness training, demonstrating improvements in psychological parameters with medium effect sizes (Cavanagh et al., 2013; Gl€uck and Maercker, 2011).
The effects of technology-supported mindfulness programs have been explored
further in human-computer interaction research. Recognizing that novices often
have trouble maintaining focus during a mindfulness session, Pisa et al. (2017) found
the act of audibly amplifying a person’s sound of breathing during meditation can
help them regain and maintain focus on their breathing, leading to a more effective
mindfulness session. A novel interactive biofeedback mobile application (PauseTM,
ustwo studio Ltd., London, England) was used by Ahmed et al. (2017) to explore the
effects of three modalities - audio, vision, touch - and their combinations on
perceived stress and the physiological indicator of relaxation measured by delta heart
rate. The Touch condition required participants to slowly draw circular movements
on the touch screen. In the Vision condition participants were asked to focus on the
dynamic changing shape of the floating bubbles. The Audio condition had the participants listen to the audio of instrumental music with nature sounds and close their
eyes whenever they felt like it. Each participant performed all conditions in a randomized crossover design and all measures in this study provided convergent results,
though the study lacked a control condition. Participants performed best in the Audio
condition and when Vision and Touch conditions were combined with Audio, participant performance in those conditions improved, suggesting that Audio is a significant contributor to overall relaxation. Conversely, Ahmed and associates (2017)
found that when Audio was combined with other senses, the positive effect was
less pronounced than when Audio was used on its own.
3
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
The studies on mindfulness effects reviewed above have employed both subjective
and objective measures of stress and relaxation. Due to the known relationship between stress and the body, there is interest in exploring more objective metrics
regarding how the body’s stress response functions in response to intervention.
Studies show consistent physiological responses across multiple methods when
attention tasks involve controlled breathing (Krygier et al., 2013; Lehrer et al.,
1999; Tang et al., 2015). Parasympathetic vagal nerve control maintains heart
rhythm in pace with breathing, also known as respiratory sinus arrhythmia (RSA;
Porges, 2007). During mindful breathing practices, RSA predominates over sympathetic control of heart rhythm, and participants demonstrate a consistent interval between heartbeats (inter-beat-interval, IBI), which corresponds with a significant
decrease in heart rate variability (HRV), with significantly greater decrease observed
in experienced practitioners (Krygier et al., 2013; Peressutti et al., 2010). Postintervention HRV is increased, which is associated with lower sympathetic activity
and positive health outcomes (Kristal-Boneh et al., 1995), this change is also more
pronounced with practice. Mindfulness states are associated with enhanced EEG
alpha (8e12 Hz) and theta (4e8 Hz) power, compared to resting state (Takahashi
et al., 2005). Generalizability and interpretation distinct EEG features, such as blocking or further decomposition into fast and slow components, is complicated by
differing methodologies in spectral analysis and variety of meditation styles examined (Ivanovski and Malhi, 2007). Acute stressors are known to increase perspiration
and skin conductance (electrodermal activity, EDA). Several studies show reduction
in EDA during mindfulness tasks, lower baseline EDA following MBSR training,
and lower EDA responsiveness to aversive stimuli (Delmonte, 1985; Grecucci
et al., 2015; Lush et al., 2009). Acute stressors increase salivary activity of digestive
enzyme a-amylase (sAA; Kirschbaum et al., 1993), which is regulated by sympathetic and parasympathetic innervation of its secretory glands, though no studies
of the sAA biomarker following a “stress reducing” condition have been published.
Increased public interest in the beneficial effects of mindfulness practices has created
an emerging market in mindfulness aids. These include mobile applications for instruction and managing practice, or electronic devices providing cues to increased
stress or mindfulness practice enhancement using psychophysiological measurements
associated with mindfulness practice and the stress response, such as respiration, heart
rate variability (HRV), electrodermal activity (EDA), or electroencephalography
(EEG). Such devices include the “Spire” (Spire Inc., San Francisco, California), which
continuously monitors movement associated with respiration to cue the user and
prompt awareness of increased stress, and products enabling cues from HRV sensors
on the wrist, such as the Apple Watch (Apple Inc., Cupertino, California). Biofeedback methodologies using electromyography, electroencephalography, or cardiac
measures have demonstrated efficacy in management of anxiety, attention deficit,
and multiple psychosomatic conditions (Wenck, et al., 1996; Kaiser and Othmer,
4
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
2000; McKee, 2008). Potential applications of these methodologies in stress management has been examined for decades (Brown, 1977).
The “Muse” is a consumer device and mobile application (InteraXon Inc., Toronto,
Canada) that aims to enhance mindfulness meditation practice, combining instruction and tracking with EEG sensor biofeedback (or neurofeedback) during mindfulness relaxation exercise. The device has been validated in a study of EEG eventrelated potentials (ERP), compared with research-grade EEG hardware (Krigolson
et al., 2017). Participants performed multiple trials of standard oddball and
reward-learning tasks, with reward positivity, N200, and P300 ERP components recorded using either the Muse device or ActiChamp 10e20 EEG system (Brain Products GmbH, Gilching, Germany). EEG system data were reduced to match the Fpz,
AF7, AF8, TP9, and TP10 channels available with the Muse device. ERP components were observed and quantifiable with both the Muse and reduced ActiChamp
EEG systems, with equivalent mean peak amplitudes. Controlled studies of EEGneurofeedback demonstrate enhanced performance in varied psychometric tests,
including sustained attention, executive function, memory, psychomotor skills,
and well-being. In the following studies, we evaluated the neurophysiological concomitants and proprietary brain activity measure of the “Muse” mindfulness device
and relaxation exercise instruction. Considering the documented performance
enhancement in diverse cognitive tasks using biofeedback methodologies, we hypothesized that the auditory EEG neurofeedback during a Muse assisted relaxation
exercise (MARE) would result in greater changes of physiological measures of
relaxation and stress compared to an unassisted relaxation exercise (URE). Specifically, this study sought to test the following research hypotheses:
Hypothesis 1: The MARE condition will result in greater changes in short-term
HRV, EDA, and sAA activity than the URE condition.
Hypothesis 2: The MARE condition will result in greater percentage of time spent in
the “Calm” mode, Muse’s EEG-derived measure of relaxation, than the URE
condition.
We also sought to answer the following exploratory research questions: Is the subjective perception of Muse audio feedback as distracting or irritating associated with
HRV and EDA changes during the MARE condition?
2. Method
2.1. Participants
Both studies were approved by the University of Florida Institutional Review Board,
and participants were recruited through an undergraduate research study
5
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
participation program and word of mouth. All participants completed the informed
consent process prior to participation. Eligible participants were between 18 and 72
years old. In Study 1, interim power analysis of Baseline and Endpoint HRV (95%
CI for difference ¼ [0.03, 0.21], n ¼ 40, d ¼ 0.37) indicated 97 participants would
need to be enrolled for sufficient power (1 - b < 0.20), therefore 99 participants were
enrolled, and 96 completed both conditions. Study 2 enrolled 46 participants to
further evaluate HRV, 23 and 21 participants completed MARE and URE with
follow-up observation.
2.2. Procedure
2.2.1. Study 1: Randomized sequential study
The randomized sequential condition study was designed to evaluate the effects of
MARE and URE relaxation conditions on HRV and EDA during relaxation, saliva
samples following each condition, and Muse application “Calm %” metric during the
relaxation. It also evaluated changes in HRV and EDA following both conditions
and effect of condition order (see Fig. 1 for further detail). MARE or URE were
administered in a counterbalanced order. In both conditions, participants engaged
in a 1-min mental word association exercise to calibrate the Muse brain activity
sensor prior to relaxation. The application provides a randomly chosen category
every 20 s, such as “tools”, “actors”, or “fruit”, and instructs the participant to think
of as many examples in that category as possible. The application then provides
auditory instructions for the relaxation condition. After both conditions were complete, follow-up HRV was recorded for 7 min while participants completed a
post-questionnaire on their perceptions of the relaxation experience.
2.2.2. Study 2: Randomized parallel study
The randomized parallel study compared the effect of a single MARE or URE on
post-intervention HRV over three observation periods following the intervention.
In either condition, participants calibrate the Muse brain activity sensor using the
same word association task as in the first study, followed by auditory instructions
for the relaxation condition (MARE or URE). Participants then engaged in the relaxation exercise for 7 min. HRV was subsequently collected for three additional 7minute intervals (see Fig. 2).
2.3. Experimental conditions
2.3.1. Muse assisted relaxation exercise (MARE)
Prior to the MARE condition, the Muse application orally explained the audio neurofeedback mechanism to participants: as more focused and relaxed brain states are
6
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Fig. 1. Initial study procedure flowchart.
identified by the device, the application provided more pleasant peaceful natural
sounds, such as wind, water, trees, and birds; conversely, as users drift out of mindful
states, turbulent and more intense sounds of wind give a clue to return their attention to
their breathing. The mindful breathing instruction tells participants to focus attention
on sensations caused by movements of the breath in the body over the course of the
relaxation exercise and to re-assert the focus as they identify changes in the audio feedback. The basic instruction is consistent with traditional breath-focused mindfulness
practices. During MARE, the Muse application uses a pair of headphones to provide
audio neurofeedback to participants based on registered brain activity. The user then
received neurofeedback through auditory cues derived from EEG activity analyzed
based on proprietary algorithms while completing the mindfulness relaxation exercise.
After the 7-minute MARE, the Muse device provided information regarding what percentage of time had been spent in a “calm” state as well as a graph depicting brain activity over time, with brain activity classified as “calm,” “neutral,” or “active.”
2.3.2. Unassisted relaxation exercise (URE)
Prior to beginning the URE condition, each participant listened to the Muse application instructions as in the MARE condition but was then instructed by the study
Fig. 2. Follow-up study procedure flowchart.
7
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
facilitator to close their eyes and “relax using whatever method you find comfortable
or intuitive”. Instead of audio neurofeedback from the Muse application, participants
listened to sham sound, a continuous recording of audio from the Muse application
when registering maximally “calm” state. The Muse application registered brain activity for the “Calm %” metric.
2.3.3. Muse “Calm” mode percentage
The Muse device uses proprietary algorithms to register and analyze brain activity
sensor data. The application then calculates the amount of time spent in the “calm”
state. The Muse manufacturer provides a software developer kit and documentation,
detailing hardware and streaming data specifications for the unmodified consumer
Muse 2014 device and version 7.8.0 firmware used in this study (InteraXon, Inc.
2018). Activity is registered in four 240hz channels using static, dry electrodes on
a flexible “headband” device. Two silver electrodes correspond to AF7 and AF8 frontal lobe positions, two conductive silicone-rubber electrodes are placed at TP9 and
TP10 temporoparietal lobe positions, and a Fpz reference electrode is located between AF7/AF8 electrodes at the prefrontal cortex midline. Additionally, a threeaxis accelerometer registers the users head movements. InteraXon provides access
and documentation to a software developer kit which allows streaming and saving
data from the Muse device, including multiple real time measures. Continuous accelerometer, real time voltage, raw fast Fourier transform, absolute power, and spectrum
power relative to total power is provided for Delta (1e4Hz), Theta (4e8Hz), Alpha
(7.5e13Hz), Beta (13e30Hz), and Gamma (30e44Hz) EEG frequency spectrums in
each channel. Other scale measures include electrode contact quality, connection
quality, blinking, and jaw clench. Experimental Boolean state measures “Concentration” and “Mellow” are sampled at 10Hz based on ratio of Gamma and Alpha powers,
with additional proprietary processing. Session “Calm” percentage value is only
available through the mobile Muse application and specific information regarding
its processing is not provided. In the researchers’ experience, a combination of the
above measures is used. Head, eye, and jaw movements during relaxation exercise
reliably result in lower Calm % values. Intentionally interfering with the wordassociation calibration step before relaxation exercise appears to shift baseline upper/lower thresholds for activity and reliably affect the subsequent session Calm %:
not performing word association and calibrating during a resting, meditative state decreases session Calm %, while an effortful disordered state results in higher Calm %.
2.3.4. Electro-dermal
(Empatica E4)
activity
and
heart
rate
variability
The Empatica E4 (Empatica Inc., Cambridge, Massachusetts) is a photoplethysmography (PPG) biosensor device placed around the participant’s wrist on the non-
8
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
dominant hand, providing similar data quality to portable clinical ECG devices
(McCarthy et al., 2016). By measuring absorption of emitted light, it calculates heart
rate, inter-beat-interval (IBI), and skin temperature, used as indices of the body’s
physiological stress and relaxation response. Time domain heart rate variability
(HRV) in the 7-minute observation intervals was calculated using the root mean
square of the successive differences (RMSSD) of IBI. The RMSSD HRV measure
captures high-frequency activity, of which parasympathetic vagal nerve activity is
a primary component (Electrophysiology TF, 1996). The E4 uses proprietary algorithms to filter PPG data artifacts and report only valid IBIs, observation intervals
containing >120 valid IBIs were used for analysis (90.6% in Study 1, 99.1% in
Study 2).
The E4 also measures electrodermal activity (EDA) using impedance from two dry
electrodes on the wrist. For 40 participants in the sequential study and all 44 participants analyzed in the parallel study, the EDA sensor was modified by attaching
external leads and adhesive paste electrodes on the middle and ring fingers of their
non-dominant hand, which provided the greater dynamic range necessary for analysis. EDA is recorded as mean impedance in microSiemens for each 7-minute
time domain interval used in HRV. Positively skewed HRV and EDA data were
normality transformed using the natural log of HRV and square root of EDA prior
to analysis.
2.3.5. Salivary a-amylase bio-assay
For collection of saliva samples in the repeated interventions study, the participant is
asked to chew and dispense a small disposable swab at 5 time points, approximately
10 min apart for conditioned samples (see Fig. 1). These sampling intervals are
consistent with Trier Social Stress Test protocols showing significant changes
following psychosocial stressors (Kirschbaum et al., 1993). Samples were analyzed
using an in vitro enzymatic activity assay of a-amylase (sAA), a physiological indicator of heightened psycho-social stress. Laboratory assay with 2mM CNP-G3 substrate was assembled using commercial components, comparable to standard
protocol (Winn-Deen et al., 1988), and validated using commercial assay kit and
standards (r ¼ 1.00, p < 0.001; RE80111, IBL International GmbH). A requirement
of the assay is that participants not eat or drink anything for 30 min prior to the study,
individual samples with undiluted sAA concentration <1U/mL were treated as
missed samples and excluded from analysis, with no imputation. Positively skewed
sAA data was natural log transformed to normality prior to analysis.
2.3.6. Perceived Stress Scale
Participants completed the Perceived Stress Scale (PSS-10, Cohen et al., 1983), a 10item self-report questionnaire that quantifies a participant’s subjective stress
9
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
experience. It is comprised of a Perceived Helplessness (PH) factor with six items,
such as “In the last month, how often have you been angered because of things that
were outside of your control?”, and a Perceived Self-Efficacy (PSE) factor with four
items, such as “In the last month, how often have you been able to control irritations
in your life?”. Responses are graded on a Likert-type scale ranging from 0 (never) to
4 (very often). PSE responses are reverse-coded and added to PH to produce the
PSS-10 score. Internal consistency of the measure reached a ¼ .86 in the current
study.
2.3.7. Post-questionnaire
Participants also completed a post-questionnaire at the end of the studies. Two
free-response questions asked for participants’ prior experience with mindfulness
or stress-reduction practice and a description of the relaxation method they implemented during the URE condition. It also included eight Likert Scale questions
with five options: “Strongly Agree”, “Agree”, Neither Agree nor Disagree”,
“Disagree”, and “Strongly Disagree”. Four Likert Scale items assessed participants’ perception of relaxation state and effort during MARE and/or URE, such
as “I felt calm or relaxed during the unassisted exercise” and “I found the
Muse audio feedback distracting or irritating.” An additional four items assessed
their interest in MBSR and mindfulness research, such as “I am interested in
the subject of mindfulness, meditation, or stress reduction” and “I would consider
using the Muse application outside of the study”. These responses allowed us to
evaluate the participants’ results based on experience with meditation or other
relaxation techniques, self-report of effort and relaxation, and interest in the study
topic.
2.4. Data analysis
E4 session data was imported and processed to mean impedance EDA and RMSSD
HRV for individually variable 7-minute observation intervals using the Python
Pandas data analysis library. Prior to analysis, HRV, sAA, and EDA data were transformed to achieve normality using natural logarithm transformation for HRV and
sAA, square root transformation for EDA.
Statistical analysis was performed using IBM SPSS 23.0 software and the R ‘gee’
package v4.19. Generalized Estimating Equations (GEE) were used for HRV,
sAA, EDA, and Calm % (in the initial study) to determine main and interaction effects of observation interval, questionnaire responses, intervention order (in the
sequential interventions study), and relaxation condition. Bonferroni corrected
post-hoc multiple comparisons and Independent-Samples Kruskal-Wallis tests
were used when significant effects were observed.
10
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
To compare Calm % measures in the parallel study, Independent-Samples MannWhitney
U
and
Independent-Samples
Kruskal-Wallis
tests
were
used.
Kolmogorov-Smirnov test was used to compare distributions of Likert Scale responses. Pearson Product-Moment and Spearman Rank-Order correlations were
used to evaluate correlations between measures collected in each study.
Neither of the experiments reported in this article was formally preregistered.
Neither the data nor the materials have been made available on a permanent thirdparty archive; requests for the data or materials can be sent via email to the corresponding author at rbussing@ufl.edu.
3. Results
3.1. Study 1: Randomized sequential condition study
3.1.1. HRV analysis
In this study, HRV was modeled as a response variable using GEE with AR(1) correlation structure, normal distribution, and identity link function (Table 1). Significant effect of interval on HRV was found, X2(3, N ¼ 362) ¼ 70.16, p < 0.001, with
no significant effect of condition order or significant interaction effect (Table 2).
Corrected multiple comparisons of intervals showed a significant decrease in HRV
in Exercise #1 vs Baseline, p ¼ 0.001, d ¼ 0.26, followed by a significant increase in
the Endpoint interval, p < 0.001, d ¼ 0.53 (Table 3). These results indicate a beneficial increase in HRV following both relaxation conditions, regardless of order of
MARE and URE conditions.
3.1.2. EDA analysis
EDA was modeled as a response variable using GEE with AR (1) correlation structure, normal distribution, and identity link function. Significant effect of interval on
EDA was found, X2(3, N ¼ 139) ¼ 95.81, p < 0.001, with no significant effect of
condition order or significant interaction effect (Table 4).
Table 1. Estimated marginal means of HRV for Study 1 by observation Interval.
Interval
Mean
Std. Error
95% Wald Confidence Interval
Lower Bound
11
Upper Bound
Baseline
4.0581
.03642
3.9868
4.1295
Exercise #1
3.9679
.03575
3.8978
4.0379
Exercise #2
4.0486
.03879
3.9725
4.1246
Endpoint
4.2372
.03742
4.1639
4.3106
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Table 2. GEE model effects of Interval and Condition Order on HRV in Study 1.
Source
Type
df
Wald Chi-Square
Sig.
(Intercept)
1
16582.608
.000***
Interval
3
70.161
.000***
Condition Order
1
2.821
.093
Interval * Condition Order
3
1.138
.768
Note: ***p < .001.
Corrected multiple comparisons of intervals showed a significant decrease in EDA in
the first relaxation condition vs baseline, p ¼ 0.01, d ¼ 0.10, followed by a significant increase in EDA in the final interval, p < 0.001, d ¼ 0.26 (Table 5). These results indicate a significant reduction in EDA initially when beginning the relaxation
conditions, with an increase once both have been completed, with no effect of intervention order.
3.1.3. Percentage of time in muse “Calm” mode
In the Randomized Sequential Condition study, Calm % was modeled using GEE with
exchangeable correlation structure, Gamma distribution, and Log link. Condition interval was used as a within-subject variable and condition order (MARE first or URE
first) as a factor. Significant effect of condition interval on Calm % was found, X2(1, N
¼ 186) ¼ 5.61, p ¼ 0.02, with no significant effect of condition order (Table 6).
Table 3. Pairwise comparisons of HRV by observation Interval with corrected
significance in Study 1.
(I)
Interval
(J)
Interval
Mean
Difference
(I-J)
Std.
Error
df
Bonferroni
Sig.
Lower
Upper
Baseline
Exercise #1
Exercise #2
Endpoint
.0902a
.0096
-.1791a
.02396
.03180
.03553
1
1
1
.001***
1.000
.000***
.0270
-.0743
-.2728
.1535
.0935
-.0854
Exercise #1
Baseline
Exercise #2
Endpoint
-.0902a
-.0807a
-.2693a
.02396
.02561
.03320
1
1
1
.001***
.010**
.000***
-.1535
-.1482
-.3569
-.0270
-.0131
-.1817
Exercise #2
Baseline
Exercise #1
Endpoint
-.0096
.0807a
-.1887a
.03180
.02561
.03720
1
1
1
1.000
.010**
.000***
-.0935
.0131
-.2868
.0743
.1482
-.0905
Endpoint
Baseline
Exercise #1
Exercise #2
.1791a
.2693a
.1887a
.03553
.03320
.03720
1
1
1
.000***
.000***
.000***
.0854
.1817
.0905
.2728
.3569
.2868
Note: (a) e Significant difference at p < .05; *p < .05; **p .01; ***p .001.
12
95% Wald
Confidence
Interval for
Difference
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Table 4. GEE model effects of Interval and Condition Order on EDA in Study 1.
Source
Type III
df
Wald Chi-Square
Sig.
(Intercept)
1
251.759
.000***
Interval
3
95.809
.000***
Condition Order
1
.843
.359
Interval * Condition Order
3
1.983
.576
Note: ***p < .001.
Table 5. Pairwise comparisons of EDA by interval with corrected significance in
Study 1.
(I)
Interval
(J)
Interval
Mean Difference
(I-J)
Std.
Error
df
Bonferroni
Sig.
95% Wald
Confidence
Interval for
Difference
Lower
Upper
Baseline
Exercise #1
Exercise #2
Endpoint
.1805a
.0600
-.4867a
.05922
.08106
.06344
1
1
1
.014*
1.000
.000***
.0243
-.1538
-.6541
.3368
.2739
-.3194
Exercise #1
Baseline
Exercise #2
Endpoint
-.1805a
-.1205
-.6673a
.05922
.06505
.08651
1
1
1
.014*
.384
.000***
-.3368
-.2921
-.8955
-.0243
.0512
-.4390
Exercise #2
Baseline
Exercise #1
Endpoint
-.0600
.1205
-.5468a
.08106
.06505
.07072
1
1
1
1.000
.384
.000***
-.2739
-.0512
-.7334
.1538
.2921
-.3602
Endpoint
Baseline
Exercise #1
Exercise #2
.4867a
.6673a
.5468a
.06344
.08651
.07072
1
1
1
.000***
.000***
.000***
.3194
.4390
.3602
.6541
.8955
.7334
Note: (a) e Significant difference at p < .05; *p < .05; ***p < .001.
Pairwise comparisons of estimated marginal means of Calm % showed significantly
higher Calm % in the second relaxation exercise than the first (95% CI ¼
[1.15,11.70], p ¼ 0.017, d ¼ 0.27), regardless of order of MARE and URE.
MARE Calm % was significantly, moderately positively correlated with URE
Calm % (rs ¼ 0.45, p < 0.001).
3.1.4. Salivary bio-assay analysis
No significant differences in sAA were found between saliva samples at the 5 time
points (Baseline: N ¼ 85, M ¼ 3.87, SD ¼ 1.01; Sample 2: N ¼ 86, M ¼ 3.96, SD ¼
1.03; Sample 3: N ¼ 86, M ¼ 4.00, SD ¼ 1.04; Sample 4: N ¼ 86, M ¼ 4.07, SD ¼
0.95; Endpoint: N ¼ 87, M ¼ 3.92, SD ¼ 1.00). Results are consistent with those
observed in unstimulated saliva samples from healthy adults (Mandel, et al., 2010).
13
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Table 6. GEE model for Muse Calm % by Condition Interval and Condition order
in Study 1.
Source
Type III
df
Wald Chi-Square
Sig.
(Intercept)
1
4543.252
.000***
Condition Interval
1
5.610
.018*
Condition Order
1
.652
.420
Condition Interval * Condition Order
1
.176
.675
Note: *p < .05; ***p < .001.
3.1.5. Perceived stress scale analysis
Participants reported overall perceived stress scores consistent with normative statistics (PSS-10 M ¼ 16.13, SD ¼ 5.61; Perceived Helplessness M ¼ 11.00, SD ¼ 4.12;
Perceived Self-Efficacy M ¼ 5.13, SD ¼ 2.20). Internal consistency was acceptable
for Helplessness (a ¼ .84), Self-Efficacy (a ¼ 0.69), and the overall PSS-10 (a ¼
0.86). No significant GEE model effects of PSS score or correlations with other measures collected in the study were found.
3.1.6. Post-questionnaire analysis
Only responses to “I found the Muse audio feedback distracting or irritating” showed
significant effects on other measures in the study, with 2 (2.1%) responding
“Strongly Agree” and 24 (25.0%) Agree”, while 34 (35.4%) responded “Disagree”
and 12 (12.5%) “Strongly Disagree”. Those who found Muse audio feedback
more distracting were lower performers in Calm % in the first relaxation condition,
though demonstrated higher HRV following both conditions. Perception of effort
during the MARE condition had significant effect when included as a factor in
GEE model of HRV, X2(4, N ¼ 362) ¼ 86.63, p < 0.001, and Calm % X2(4, N
¼ 186) ¼ 18.80, p < 0.001, with no significant interaction effects with condition
or order of MARE/URE. Corrected multiple comparisons demonstrated statistically
significant difference between HRV in follow-up observation interval groups responding “Agree” (n ¼ 20) and “Neither agree nor disagree” (n ¼ 23), (95% CI
¼ [0.02,0.57], p ¼ 0.027, d ¼ .86), and “Agree” and “Strongly Disagree” (n ¼
11), (95% CI ¼ [0.04, 0.71], p ¼ 0.02, d ¼ 1.15). Corrected non-parametric multiple
comparisons of Calm % demonstrated significant difference in Calm % performance
between those responding “Agree” and “Strongly Disagree” (Z ¼ -2.77, p ¼ 0.03) in
the first relaxation condition only. No other significant effects of Post-Questionnaire
responses on measures collected in this study were found.
14
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
3.2. Study 2: Randomized parallel conditions
3.2.1. HRV analysis
A similar effect of interval on GEE model of HRV was identified in the parallel
study. Significant effect of interval on HRV was found, X2(4, N ¼ 218) ¼ 54.71,
p < 0.001, with no significant effect of Condition or significant interaction effect.
Corrected multiple comparisons of intervals showed a significant decrease in
HRV during the relaxation condition vs baseline, 95% CI ¼ [-0.11-0.22], t(43) ¼
6.19, p < 0.001, d ¼ 0.42. Similar to Study 1, HRV results in Study 2 indicate a
beneficial increase in HRV in the first follow-up observation period vs baseline
(95% CI ¼ [0.00, 0.16], t(43) ¼ 2.03, p ¼ 0.049, d ¼ 0.22), with no difference between MARE and URE or in additional observation periods.
3.2.2. EDA analysis
Significant effect of interval on GEE model of EDA was found, X2(4, N ¼ 203) ¼
47.495, p < 0.001, and a significant interaction effect between interval and condition, X2(4, N ¼ 203) ¼ 12.86, p ¼ 0.012. Condition remains an insignificant factor
in EDA in GEE model without interaction effect. Multiple comparisons in Study 2
display the same overall trend of increase in skin conductivity in follow-up observation, unrelated to the study conditions.
4. Discussion
4.1. Effect of neurofeedback on stress reduction
These study findings did not support the initial hypothesis that the EEG-based neurofeedback during the MARE condition would result in greater changes of physiological measures of relaxation and stress compared to the URE condition. Instead, the
MARE and URE conditions were found to be equally effective in producing decreases in HRV during the exercises, followed by beneficial short-term increases
in HRV that are associated with stress reduction. Observed changes in HRV are
consistent with findings in prior mindfulness practice studies (e.g., Krygier et al.,
2013; Peressutti et al., 2010).
Calm percentage was increased in the second relaxation exercise, regardless of the
order of the MARE and URE conditions. Changing MARE and URE conditions
had no effect on Calm percentage, and no effect of self-reported prior experience
was found. MARE Calm percentage was significantly, moderately positively correlated with the URE Calm percentage (rs ¼ 0.45, p < 0.001). Interestingly, results of
the post-intervention perceptions questionnaire revealed that those who agreed that
the Muse audio feedback was “distracting or irritating” had substantially lower Calm
% performance, supporting it as a measure of participant effort during the relaxation
15
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
exercise. However, this was only in the first relaxation condition, and the same group
of participants also demonstrated higher HRV following the mindfulness exercise. It
appears that the negative perception of Muse audio feedback in this subset of participants may have led to greater difficulty in initial engagement with the mindfulness
exercise. The substantially larger HRV effects in participants who found it more distracting suggests that higher effort relaxation may produce more benefit.
Furthermore, against our initial hypothesis, no difference was found between MARE
and URE conditions in the effects of relaxation on EDA. Several studies have
demonstrated reduction in EDA during mindfulness exercises, lower baseline
EDA following MBSR training, and lower EDA responsiveness to aversive stimuli
(Delmonte, 1985; Erisman and Roemer, 2010; Grecucci et al., 2015; Lush et al.,
2009). Reports on this measure are less consistent compared to the effects of mindfulness on HRV (Erisman and Roemer, 2010) and our sequence of two studies indicate that while a significant reduction in EDA was initially achieved by participants
when beginning relaxation conditions, there was a significant increase once both interventions had been completed, with no effect of intervention order. This finding is
consistent with the HRV results discussed above, suggesting that stress reduction is
indeed achieved due to sustained focus and relaxation during mindfulness exercise,
but this effect is not statistically different across the two experimental conditions and
therefore shows no additional benefit of the EEG neurofeedback in this specific implementation of the technology.
Finally, another measure used in our studies to explore physiological changes in
stress levels was salivary alpha amylase (sAA) activity. SAA is increased following
stress inducing conditions (Kirschbaum et al., 1993), though our sequence of experiments failed to show a significant difference in sAA activity in response to either
MARE or URE. This suggests that unlike HRV and EDA, sAA activity may not
be sensitive enough as a physiological measure of stress to detect stress level
changes, especially reductions in response to brief relaxation exercises.
4.2. Research contributions
Overall, the results of our study are consistent with results of prior research on
mindfulness-based relaxation and its positive effects on stress reduction (Carmody
and Baer, 2009). Importantly, even brief mindfulness-based interventions, such as
the Muse-assisted and unassisted relaxation exercises, may offer short-term benefits
in the form of increased HRV. Our findings contribute to the growing body of literature on technology-enabled relaxation approaches (Cavanagh et al., 2013; Gl€uck
and Maercker, 2011; Krusche et al., 2012) suggesting that while short-term stress
reduction may indeed be achieved by engaging in simple, brief relaxation exercises,
such benefits cab be achieved without externalizing or “outsourcing” primary cognitive awareness components of meditation and mindful awareness practices.
16
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
4.3. Research limitations
Limitations of this study include the single participation sessions, hardware, and analysis. A longitudinal study may observe persistent improvements in stress measures or
training effects associated with neurofeedback that this study cannot detect (Zoefel
et al., 2011). While simple to use, the Empatica E4 sensor limits quality and detailed
spectral analysis of physiological signals compared to ECG methods, and analysis
was limited to more accessible mean EDA and RMSSD HRV time domain measures.
Future studies may explore the unexpected finding that increased difficulty in initially
engaging attentional awareness may affect more positive physiological outcomes.
5. Conclusion
A growing body of literature shows that mindfulness-based exercises improve regulation of emotion and cognition. Biofeedback and other technologies have been
recently introduced to assist individuals with mindfulness, mediation, and stress
reduction by externalizing components of mindfulness practice. Our investigation
of the effects of one such device e Interaxon’s Muse e has failed to show incremental short-term stress reduction benefits above unassisted relaxation, using the physiological indices of heart rate variability, electro-dermal activity, and Muse’s own
EEG-based measure, percentage of time spent in Calm mode. Future studies should
explore if longer, repeated mindfulness interventions with Muse or other devices
would produce more pronounced incremental effects of EEG neurofeedback on
stress reduction and well-being.
Declarations
Author contribution statement
Artem S. Svetlov: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.
Melanie M. Nelson, Joseph P.H. McNamara: Conceived and designed the experiments; Wrote the paper.
Pavlo D. Antonenko: Analyzed and interpreted the data; Wrote the paper.
Regina Bussing: Conceived and designed the experiments; Analyzed and interpreted
the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Funding statement
This research did not receive any specific grant from funding agencies in the public,
commercial, or not-for-profit sectors.
17
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Competing interest statement
The authors declare no conflict of interest.
Additional information
No additional information is available for this paper.
References
Ahmed, M.M.H., Silpasuwanchai, C., Niksirat, K.S., Ren, X., 2017. Understanding
the role of human senses in Interactive meditation. In: Proceedings of the 2017 CHI
Conference on Human Factors in Computing Systems (CHI ’17). ACM, New York,
NY, pp. 4960e4965.
American Institute of Stress, 2013. Workplace Stress. http://www.stress.org/
workplace-stress/.
American Psychological Association, 2017. Stress in Amercia: the State of Our Nation.
Retrieved from. http://www.apa.org/news/press/releases/stress/2017/state-nation.pdf.
Bodhi, Bhikkhu, 2011. What does mindfulness really mean? A canonical perspective. Contemp. Buddhism 12 (1), 19e39.
Brosschot, J.F., Pieper, S., Thayer, J.F., 2005. Expanding stress theory: prolonged
activation and perseverative cognition. Psychoneuroendocrinology 30 (10),
1043e1049.
Brown, B.B., 1977. Stress and the Art of Biofeedback. Harper & Row.
Carlson, L.E., Speca, M., Patel, K.D., Goodey, E., 2004. Mindfulness-based stress
reduction in relation to quality of life, mood, symptoms of stress and levels of
cortisol, dehydroepiandrosterone sulfate (DHEAS) and melatonin in breast and
prostate cancer outpatients. Psychoneuroendocrinology 29 (4), 448e474.
Carmody, J., Baer, R.A., 2008. Relationships between mindfulness practice and
levels of mindfulness, medical and psychological symptoms and well-being in a
mindfulness-based stress reduction program. J. Behav. Med. 31 (1), 23e33.
Carmody, J., Baer, R.A., 2009. How long does a mindfulness-based stress reduction
program need to be? A review of class contact hours and effect sizes for psychological distress. J. Clin. Psychol. 65 (6), 627e638.
Cavanagh, K., Strauss, C., Cicconi, F., Griffiths, N., Wyper, A., Jones, F., 2013. A
randomised controlled trial of a brief online mindfulness-based intervention. Behav.
Res. Ther. 51 (9), 573e578.
18
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Cohen, S., Kamarck, T., Mermelstein, R., 1983. A global measure of perceived
stress. J. Health Soc. Behav. 24 (4), 385e396.
Collard, P., Avny, N., Boniwell, I., 2008. Teaching mindfulness based cognitive
therapy (MBCT) to students: the effects of MBCT on the levels of mindfulness
and subjective well-being. Counsell. Psychol. Q. 21 (4), 323e336.
Delmonte, M.M., 1985. Effects of expectancy on physiological responsivity in
novice meditators. Biol. Psychol. 21 (2), 107e121.
DeVibe, M., Bjørndal, A., Tipton, E., Hammerstrøm, K.T., Kowalski, K., 2012.
Mindfulness based stress reduction (MBSR) for improving health, quality of life,
and social functioning in adults. Campbell Sys. Rev. 8 (3).
Del Re, A.C., Fl€
uckiger, C., Goldberg, S.B., Hoyt, W.T., 2013. Monitoring mindfulness practice quality: an important consideration in mindfulness practice. Psychother. Res. 23 (1), 54e66.
Dobkin, P.L., Zhao, Q., 2011. Increased mindfulnesseThe active component of the
mindfulness-based stress reduction program? Complement. Ther. Clin. Pract. 17
(1), 22e27.
Erisman, S.M., Roemer, L., 2010. A preliminary investigation of the effects of
experimentally induced mindfulness on emotional responding to film clips.
Emotion 10 (1), 72e82.
Farb, N.A.S., Segal, Z.V., Anderson, A.K., 2013. Mindfulness meditation training
alters cortical representations of interoceptive attention. Soc. Cognit. Affect Neurosci. 8 (1), 15e26.
Gl€
uck, T.M., Maercker, A., 2011. A randomized controlled pilot study of a brief
web-based mindfulness training. BMC Psychiatry 11 (1), 175.
Goyal, M., Singh, S., Sibinga, E.M., Gould, N.F., Rowland-Seymour, A.,
Sharma, R., Berger, Z., Sleicher, D., Maron, D.D., Shihab, H.M.,
Ranasinghe, P.D., Linn, S., Saha, S., Bass, E.B., Haythornthwaite, J.A., 2014.
Meditation programs for psychological stress and well-being: a systematic review
and meta-analysis. JAMA Int. Med. 174 (3), 357e368.
Grecucci, A., De Pisapia, N., Kusalagnana Thero, D., Paladino, M.P., Venuti, P.,
Job, R., 2015. Baseline and strategic effects behind mindful emotion regulation:
behavioral and physiological investigation. PLoS One 10 (1), e0116541.
Hopkins, A., Proeve, M., 2013. Teaching mindfulness-based cognitive therapy to
trainee psychologists: qualitative and quantitative effects. Counsell. Psychol. Q.
26 (2), 115e130.
19
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
InteraXon Inc, 2018. Muse: the Brain Sensing Headband Developer Portal. http://
developer.choosemuse.com/.
Ivanovski, B., Malhi, G.S., 2007. The psychological and neurophysiological concomitants of mindfulness forms of meditation. Acta Neuropsychiatr. 19 (2), 76e91.
Kabat-Zinn, J., 2003. Mindfulness-based interventions in context: past, present, and
future. Clin. Psychol. Sci. Pract. 10, 144e156.
Kabat-Zinn, J., 2011. Some reflections on the origins of MBSR, skillful means, and
the trouble with maps. Contemp. Buddhism 12, 281e306.
Kaiser, D.A., Othmer, S., 2000. Effect of neurofeedback on variables of attention in
a large multi-center trial. J. Neurother. 4, 5e15.
Kirschbaum, C., Pirke, K.M., Hellhammer, D.H., 1993. The ’Trier Social Stress
Test’–a tool for investigating psychobiological stress responses in a laboratory
setting. Neuropsychobiology 28 (1-2), 76e81.
Krigolson, O.E., Williams, C.C., Norton, A., Hassall, C.D., Colino, F.L., 2017.
Choosing MUSE: validation of a low-cost, portable EEG system for ERP research.
Front. Neurosci. 11, 109.
Kristal-Boneh, E., Raifel, M., Froom, P., Ribak, J., 1995. Heart rate variability in
health and disease. Scand. J. Work. Environ. Health 21 (2), 85e95.
Krusche, A., Cyhlarova, E., King, S., Williams, J.M.G., 2012. Mindfulness online:
a preliminary evaluation of the feasibility of a web-based mindfulness course and
the impact on stress. BMJ Open 2 (3), e000803.
Krygier, J.R., Heathers, J.A.J., Shahrestani, S., Abbott, M., Gross, J.J., Kemp, A.H.,
2013. Mindfulness meditation, well-being, and heart rate variability: a preliminary
investigation into the impact of intensive Vipassana meditation. Int. J. Psychophysiol. 89 (3), 305e313.
Lehrer, P., Sasaki, Y., Saito, Y., 1999. Zazen and cardiac variability. Psychosom.
Med. 61 (6), 812e821.
Lush, E., Salmon, P., Floyd, A., Studts, J.L., Weissbecker, I., Sephton, S.E., 2009.
Mindfulness meditation for symptom reduction in fibromyalgia: psychophysiological correlates. J. Clin. Psychol. Med. Settings 16 (2), 200e207.
MacCoon, D.G., Imel, Z.E., Rosenkranz, M.A., Sheftel, J.G., Weng, H.Y.,
Sullivan, J.C., Lutz, A., 2012. The validation of an active control intervention for
Mindfulness Based Stress Reduction (MBSR). Behav. Res. Ther. 50 (1), 3e12.
20
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Mandel, A.L., des Gachons, C.P., Plank, K.L., Alarcon, S., Breslin, P.A., 2010. Individual differences in AMY1 gene copy number, salivary a-amylase levels, and
the perception of oral starch. PLoS One 5 (10), e13352.
McCarthy, C., Pradhan, N., Redpath, C., Adler, A., 2016. Validation of the Empatica E4 wristband. In: Paper presented at the Student Conference (ISC), 2016 IEEE
EMBS International.
McKee, M.G., 2008. Biofeedback: an overview in the context of heart-brain medicine. Clevel. Clin. J. Med. 75, S31.
Peressutti, C., Martín-Gonzalez, J.M., García-Manso, J.M., Mesa, D., 2010. Heart
rate dynamics in different levels of Zen meditation. Int. J. Cardiol. 145 (1), 142e146.
Pisa, A., Nassau, A., Chernyshov, G., Junze, K., 2017. Towards interactive mindfulness training using breathing based feedback. In: Proceedings of the 2017 ACM
International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers (UbiComp ’17). ACM, New York, NY, pp. 688e692.
Porges, S.W., 2007. The polyvagal perspective. Biol. Psychol. 74 (2), 116e143.
Rosenzweig, S., Greeson, J.M., Reibel, D.K., Green, J.S., Jasser, S.A., Beasley, D.,
2010. Mindfulness-based stress reduction for chronic pain conditions: variation in
treatment outcomes and role of home meditation practice. J. Psychosom. Res. 68
(1), 29e36.
Salleh, M.R., 2008. Life event, stress and illness. Malays. J. Med. Sci. 15 (4), 9e18.
Shapiro, S.L., Jazaieri, H., Goldin, P.R., 2012. Mindfulness-based stress reduction
effects on moral reasoning and decision making. J. Posit. Psychol. 7 (6), 504e515.
Takahashi, T., Murata, T., Hamada, T., Omori, M., Kosaka, H., Kikuchi, M.,
Yoshida, H., Wada, Y., 2005. Changes in EEG and autonomic nervous activity during meditation and their association with personality traits. Int. J. Psychophysiol. 55
(2), 199e207.
Tang, Y.-Y., H€
olzel, B.K., Posner, M.I., 2015. The neuroscience of mindfulness
meditation. Nat. Rev. Neurosci. 16 (4), 213.
Task Force of the European Society of Cardiology and the North American Society
of Pacing and Electrophysiology, 1996. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Circulation 93 (5), 1043e1065.
Wenck, L.S., Leu, P.W., D’Amato, R.C., 1996. Evaluating the efficacy of a
biofeedback intervention to reduce children’s anxiety. J. Clin. Psychol. 52,
469e473.
21
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Article Nowe01351
Winn-Deen, E.S., David, H., Sigler, G., Chavez, R., 1988. Development of a direct
assay for alpha-amylase. Clin. Chem. 34 (10), 2005e2008.
Zoefel, B., Huster, R.J., Herrmann, C.S., 2011. Neurofeedback training of the upper
alpha frequency band in EEG improves cognitive performance. Neuroimage 54 (2),
1427e1431.
22
https://doi.org/10.1016/j.heliyon.2019.e01351
2405-8440/Ó 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).